四种自相关模式控制图的比较研究  

Comparative Study on Four Kinds of Autocorrelation Pattern Control Charts

作　　者：孙小素[1] Sun Xiaosu(Department of Statistics, Shandong Institute of Business and Technology, Yantai Shandong 264005, Chin)

机构地区：[1]山东工商学院统计系

出　　处：《统计与决策》2018年第9期19-23,共5页Statistics & Decision

摘　　要：学者们构造出4种自相关过程控制图。从侦测过程异常变动的角度看,这4种控制图表现并不相同。尽管已有学者对其中的某两种控制图的侦测效果进行过比较研究,但还缺乏4种控制图的整体性评价。文章选择失控过程平均链长ARL1这一指标,通过模拟方法,针对4种控制图在监控自相关过程上的表现做了详尽的比较分析。结果表明,总体来看,ARMA控制图最为灵敏,为质量管理人员的正确选择提供了理论依据。So far four different kinds of control charts have been proposed by researchers. From the perspective of detection process abnormal change, each control chart has different performance. Some researchers compare the performance of the two con- trol charts, but the integrative assessment of the four kinds is lacking. This paper takes the average run length of runaway process ARL1 as an indicator to make a detailed comparative analysis on the performance of the four control charts in monitoring autocor- relation. The results show that overall, ARMA control charts are the most sensitive and provide a theoretical basis for the correct selection of quality controllers.

关 键 词：控制图 自相关过程 比较 平均链长ARL 

分 类 号：O212[理学—概率论与数理统计]